Abstract
Suppression effect in multiple regression analysis may be more common in research than what is currently recognized. We have reviewed several literatures of interest which treats the concept and types of suppressor variables. Also, we have highlighted systematic ways to identify suppression effect in multiple regressions using statistics such as: R2, sum of squares, regression weight and comparing zero-order correlations with Variance Inflation Factor (VIF) respectively. We also establish that suppression effect is a function of multicollinearity; however, a suppressor variable should only be allowed in a regression analysis if its VIF is less than five (5).
Highlights
When selecting a set of study variables, researchers frequently test correlations between the outcome variables and theoretically relevant predictor variables [1]
We undertook a review of science literatures and various databases to understand the concept of multicollinearity and suppressor variables in regression analysis, again we went ahead to further examine the linkage between multicollinearity and suppression effect keeping in mind the supposed implication of multicollinearity in over or underestimating regression inferences
We intend to show the limitation of stepwise selection and the advantage of multicollinearity in regression analysis by evaluating the regressor weights and the general predictability of the regression model with variance inflation factor (VIF) as a constraint
Summary
When selecting a set of study variables, researchers frequently test correlations between the outcome variables (i.e. dependent variables) and theoretically relevant predictor variables (i.e. independent variables) [1]. One or more of the predictor variables are uncorrelated with the outcome variable [2] This situation poses the question of whether researchers’ multiple regression analysis should exclude independent variables that are not significantly correlated with the dependent variable [3]. Questions such as this are most times not given the supposed credit. A suppressor variable correlates significantly with other independent variables, and accounts for or suppresses some outcome-irrelevant variations in such predictors as well as improving the overall predictive power of the model. Some prefer to call the suppressor variable an enhancer [5]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
